The justice system projects the image of stoic fairness. But data doesn’t lie: it’s full of unchecked bias.

For example: judges hand out harsher sentences when they’re hungry. Or when their local football team loses. In some circuits, a claimant’s odds of winning a case can vary from 5% to 88%, depending on which individual judge is assigned to the case. The core question: why even bother with the spectacle of democratic law-making if it’s the identity of the individual judge, or factors extraneous to the law, that mostly determine the outcome? I’ve long believed that AI tools can help us identify and decrease the excessive impact of human bias in the legal system. There are some well-founded fears, but the use of advanced statistics and bigger data sets could lead to the next revolution in strengthening the rule of law. I had the privilege of writing a piece on this for The Atlantic, which you can check out here. I’ve already received some very interesting feedback on it – including from one Supreme Court judge, who was very friendly to this argument. Stay tuned, as this will be an ongoing emerging story for the legal system … we assume the fully human system we have today is the baseline, and that it has an acceptable level of fairness, but that’s only because we don’t know the alternatives yet …